Discovery of quinone-directed antitumor agents selectively bioactivated by NQO1 over CPR with improved safety profile

Article abstract of DOI:10.1016/j.ejmech.2017.02.004

In this work, we mainly focused on discovering compounds with good selectivity for NQO1 over CPR. The NQO1-mediated two-electron reduction of compounds would kill cancer cells selectively, while CPR-mediated one-electron reduction would induce potential hepatotoxicity. Several novel quinone-directed antitumor agents were discovered as specific NQO1 substrates through structure-activity relationship studies. Among them, compound 3,7,8-trimethylnaphtho[1,2-b]furan-4,5-dione (12b) emerged as the most specific substrate of the two-electron oxidoreductase NQO1 and could hardly be reduced by CPR. It afforded the highest selectivity between NQO1/CPR (selectivity ratio = 6.37), much higher than the control β-lapachone (selectivity ratio = 1.36), indicated 12b may possess superior safety profile. The electrochemical studies provided a reasonable explanation to the good selectivity toward NQO1. Molecular docking studies supported that 12b was capable of forming additional C-H … π interactions with Trp105 and Phe178 residues compared to the control β-lap. In addition, compound 12b was shown to kill cancer cells efficiently both in vitro and in vivo model. This work gave us a promising and novel scaffold for further investigation.

Full text of DOI:10.1016/j.ejmech.2017.02.004

Accepted Manuscript
Discovery of quinone-directed antitumor agents selectively bioactivated by NQO1
over CPR with improved safety profile
Jinlei Bian, Xiang Li, Nan Wang, Xingsen Wu, Qidong You, Xiaojin Zhang
PII:
S0223-5234(17)30069-7
10.1016/j.ejmech.2017.02.004
EJMECH 9200
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 25 November 2016
Revised Date: 3 February 2017
Accepted Date: 4 February 2017
Please cite this article as: J. Bian, X. Li, N. Wang, X. Wu, Q. You, X. Zhang, Discovery of quinone-
directed antitumor agents selectively bioactivated by NQO1 over CPR with improved safety profile,
European Journal of Medicinal Chemistry (2017), doi: 10.1016/j.ejmech.2017.02.004.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Discovery of quinone-directed antitumor agents selectively
bioactivated by NQO1 over CPR with improved safety profile
a
a
Jinlei Bian a,1, Xiang Li a,1, Nan Wang , Xingsen Wu a,b, Qidong You ,* Xiaojin
Zhang a,b,
*
a
State Key Laboratory of Natural Medicines, and Jiangsu Key Laboratory of Drug Design and Optimization,
China Pharmaceutical University, Nanjing 210009, China
b Department of Organic Chemistry, School of Science, China Pharmaceutical University, Nanjing 210009, China
1 These two authors contributed equally to this work.
* Corresponding authors. E-mail: youqd@163.com (Q. You), zxj@cpu.edu.cn (X. Zhang)
Abstract
In this work, we mainly focused on discovering compounds with good selectivity for
NQO1 over CPR. The NQO1-mediated two-electron reduction of compounds would
kill cancer cells selectively, while CPR-mediated one-electron reduction would induce
potential hepatotoxicity. Several novel quinone-directed antitumor agents were
discovered as specific NQO1 substrates through structure-activity relationship studies.
Among them, compound 3,7,8-trimethylnaphtho[1,2-b]furan-4,5-dione (12b) emerged
as the most specific substrate of the two-electron oxidoreductase NQO1 and could
hardly be reduced by CPR. It afforded the highest selectivity between NQO1/CPR
(selectivity ratio = 6.37), much higher than the control β-lapachone (selectivity ratio =
1.36), indicated 12b may possess superior safety profile. The electrochemical studies
provided a reasonable explanation to the good selectivity toward NQO1. Molecular
docking studies supported that 12b was capable of forming additional C-H…π
interactions with Trp105 and Phe178 residues compared to the control β-lap. In
addition, compound 12b was shown to kill cancer cells efficiently both in vitro and in
vivo model. This work gave us a promising and novel scaffold for further
investigation.
Keywords
NQO1 substrates, Ortho-quinones, ROS, β-Lapachone, Cancer cells
Abbreviations
ROS, reactive oxygen species; NQO1, NAD(P)H:quinone oxidoreductase 1; β-lap,
β-lapachone; CPR, NADPH-cytochrome P450 reductase; IBX, 2-iodoxybenzoic acid;
DDQ, 2,3-dichloro-5,6-dicyano-1,4-benzoquinone; DNQ, deoxynyboquinone;
DMSO,
dried
dimethyl
sulfoxide;
Bu4NPF6,
tetrabutylammonium
hexafluorophosphate; DCFH-DA, 2,7-dichlorodihydrofluorescein diacetate; DIC,
dicoumarol; NAC, N-acetylcysteine.
1

ACCEPTED MANUSCRIPT
1. Introduction
Most human cancers are characterized by higher levels of oxidative stress compared
with normal cells. There has been much attention directed towards the exploring
redox stress in cancers for developing new therapeutic strategies and to preferentially
kill the tumor cells [1]. Oxidative stress is developed by reactive oxygen species
(ROS) that accumulate as a result of an imbalance between ROS generation and
elimination [2]. Higher basal ROS level makes cancer cells more susceptible to
exogenously induced ROS-mediated cell death [3]. Excessive ROS levels irreversibly
damage DNA and lipids which would ultimately lead to apoptosis of cancer cells [2].
Actually, even slight increase of oxidative stress in cancer cells have been postulated
to make them vulnerable to apoptotic stimuli and exacerbate the cytotoxicity by
cancer drugs [2]. Potentiating ROS in cancers employing small molecules has been
regarded as a rational drug design strategy [3-6]. Many drugs exert their anticancer
activity dependent on the high basal ROS levels uniquely present in cancer cells [7-9].
For several ROS-enhancing compounds, including phenethylisothiocyanate (PEITC)
[10], parthenolide [11], piperlongumine [6], erastin [12], and lanperisone [13], have
been demonstrated to be selectivity for cancer cells over normal cell types in in vitro
or in vivo models of cancer.
In this regard, β-lapachone (β-lap, 1), a natural tetrahydropyran-fused
ortho-naphthoquinone isolated from the Bignoniaceae family, has gained much
attention for its efficient antitumor activity. Several studies have shown that β-lap
exerts its antitumor activity through NQO1-mediated ROS production by redox
cycling [14-16]. NQO1 is a two-electron oxidoreductase which is constitutively
overexpressed 5- to 100-fold greater than in correlative normal tissues [14]. Thus,
β-lap-mediated cell death occurs specifically in cancer cells, while normal cells with
low endogenous levels of the enzyme are spared [15]. However, its NQO1 specific
activity can only be realized within a relatively narrow concentration range and
exposure time window limited the application of β-lap [16]. The relatively narrow
therapeutic window and potential toxicity of β-lap were reported to be related to its
enzymatic one-electron reduction/activation in the presence of suitable electron
2

ACCEPTED MANUSCRIPT
donors [16]. The one-electron reduction of quinones to semiquinone species is
typically performed by the one-electron oxidoreductases (e.g., NADPH-cytochrome
P450 reductase (CPR) [17]. Given that CPR is highly expressed in human liver tissues
and CPR-mediated quinone hepatotoxicity was reported previously [18], we
hypothesized that the potential hepatotoxicity of β-lap may be caused by the reduction
of CPR (Fig. 1A). Our group has been committed to develop β-lap derivatives as
novel effective antitumor drugs, some concerns were raised to our previous work. As
shown in Fig. 1B, derivatives of β-lap (2 and 3) have been designed and synthesized
as novel NQO1-directed antitumor ortho-quinones with improved chemical stability
or enzyme catalytic efficiency [19]. However, similar to β-lap, compounds 2 and 3
could also be triggered by CPR, demonstrating the potentiality to the resultant
NQO1-irrelevant cytotoxicity of these compounds at high concentration and
prolonged treatment. Compounds with better selectivity between NQO1/CPR may
indicate superior safety profile [17, 20-21].
Structure-based ligand design is one of efficient methods in modern drug
development for targets with resolved three-dimensional structures [22-24]. Therefore,
we first investigated the catalytic pocket of NQO1 and CPR to get deeper insight of
the difference between them. The investigation of the catalytic site of NQO1 has
already been published previously [19], but the crystal complex of CPR has not yet
been reported by now. Through analyzing the native structures of CPR, we found that
the catalytic pocket is very flexible (Fig. S1). A large conformational change in CPR
is expected to take place for smooth interaction between CPR and its redox partners
and for swift progression of enzymatic reactions. The flexible property of CPR
limited us for searching efficient antitumor quinones with better NQO1/CPR
selectivity through structure-based ligand design. Thus, in this work, we developed
quinone-directed agents through the traditional medicinal chemistry efforts focus on
the selectivity between NQO1/CPR, culminating in the discovery of the promising
scaffold 12b.
3

ACCEPTED MANUSCRIPT
Fig. 1. (A) The reduction of quinone substrates by two-electron oxidoreductase NQO1 can induce
the apoptosis in tumor cells through accumulating ROS in cancer cells. The reduction of quinone
substrates by one-electron oxidoreductase CPR, which is highly expressed in human liver tissues,
can induce hepatotoxicity. (B) Structures of β-lap and our previous reported representative
compounds. The data of these compounds’ reduction rates by NQO1or CPR together with the
selective ratio are listed. The values represents the reduction rates, and the units were expressed as
µmol NADPH oxidized/min/µmol protein. The selective ratio means reduction rates by
NQO1/reduction rates by CPR.
2. Chemistry
The designed diverse range of scaffolds is shown in Fig. 2. First, substituted
furane-ortho-naphthoquinones (12a-k) is outlined in Scheme 1. Starting from
1,4-benzoquinone (6) and enamine (8), intermediate 11 was prepared via four steps
according to our previously reported methods [19]. Briefly, treatment of
1,4-benzoquinone 6 with enamine 8, prepared from aldehyde 7 and morpholine,
4

ACCEPTED MANUSCRIPT
provided the adduct product 9. Compound 10 was obtained by heating intermediate 9
in the presence of diluted hydrochloric acid. The oxidation of 10 using the reagent
2-iodoxybenzoic acid (IBX) provided the intermediate quinone 11. Then, the target
compounds 12a-k were prepared by reacting intermediate 11 and substituted
butadiene in the presence of 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ)
according to Scheme 1.
Scheme 2 describes the synthesis of dihydrofuran-fused ortho-naphthoquinones
(17a-d) and para-naphthoquinones (18 and 19). The synthetic method was according
to the procedures already published [25]. These compounds were synthesized in two
or three steps from lawsone (13). The substituted allyl bromide reacted with 13
resulted in the substituted 1,4-naphthoquinones (14 and 15) (Scheme 2). Claisen
rearrangement of 14 gave substituted allyl lawsone (16), which was then cyclized to
the ortho-quinone (17a-d) mediated by Lewis acid NbCl5 at room temperature.
Compound 15 or 16 was cyclized to the corresponding para-naphthoquinone (18 or
o
19) mediated by hydrochloric acid at 45 C. Other para-quinones (27 and 29) were
obtained based on the palladium-catalyzed C−H bond activation/C−C bond
formation/intramolecular Tsuji−Trost cascade reaction method (Scheme 3) [26]. The
key intermediate 26 to the target compound 27 was synthesized through a route
consisting of acetoacetylation of N-methyl-2,5-dimethoxyaniline (21) by
2,2,6-trimethyl-4H-1,3-dioxin-4-one, followed by Knorr cyclization of the
acetoacetamide (22) to the carbostyril (23) [27]. Oxidation and epoxidation of 23
formed epoxide 25. Finally, the key intermediate 26 was obtained through alcoholysis
of cyclohexene oxide (Scheme 3).
5

ACCEPTED MANUSCRIPT
Scheme 1. Reagents and conditions: (i) morpholine, K2CO3, 0 oC to rt, 3 h, 55-83%; (ii) CH2Cl2, 0
o
oC to rt, 12 h; (iii) dil. HCl, 75 C, 3 h, 40-62% over two steps; (iv) IBX, CH3CN/H2O, rt, 3 h,
82-90%; (v) substituted-butadiene, toluene, DDQ, 110 oC, 2-6 h, 22-35%.
Scheme 2. Reagents and conditions: (i) substituted allyl bromide, K2CO3, DMF, rt to 45 oC,
o
57-72%; (ii) toluene, 120 C, 3-6 h, 63-90%; (iii) NbCl5, DCM, rt, 0.5-1 h, 70-85%; (iv)
hydrochloric acid, 70 oC, 3 h, 73-80%.
6

ACCEPTED MANUSCRIPT
o
Scheme 3. Reagents and conditions: (i) (HCHO)n, CH3ONa, CH3OH, rt, 4 h; (ii) NaHB4, 60 C,
50% over two steps; (iii) 2,2,6-trimethyl-l,3-dioxin-4-one, xylene, 120 oC, 2 h, 75%; (iv) PPA, 95
oC, 3 h, 85%; (v) CAN, acetonitrile/H2O, rt, 45 min, 90%; (vi) H2O2, Na2CO3, ethanol, rt, 10 min,
53%; (vii) H2SO4, rt, 30 min, 44%. (viii) 28, Pd(OAc)2, Cu(OAc)2, NaOAc, THF, 45 oC, 3 h, 45%.
(ix) Pd(OAc)2, Cu(OAc)2, NaOAc, THF, 45 oC, 3 h, 82%.
3. Results and Discussion
3.1. The designation and evaluation of novel quinone-containing scaffolds.
In order to discover novel quinone compounds with high specificity towards
NQO1, a series of β-lap-inspired scaffolds were designed according to chemical
diversity and synthetic possibility (Fig. 2). Firstly, the A-ring phenyl of β-lap was
substituted by various methyl substituents at different positions. Secondly, the
ortho-quinone pharmacophore (B ring) was changed into the para-quinone scaffold. It
was worth noting that compound 27 with nitrogen-containing heteroaromatic ring was
designed as the merged molecule from two natural products
α-xiloidone and
deoxynyboquinone (DNQ) [26,28]. Thirdly, the C-ring was converted into substituted
furan ring in combination with replacing AB-ring. Finally, nine types of quinone
scaffolds were designed and synthesized. Among them, nine compounds (12b-12j)
were novel quinone scaffolds and others have been reported by us in previous work.
However, all of these quinones have not been evaluated and discussed for their
selectivity between NQO1/CPR systematically.
7

ACCEPTED MANUSCRIPT
Fig. 2. Diverse β-lap-inspired quinone scaffolds.
3.2. Determination of quinone selectivity toward NQO1 compared with CPR.
Compounds with better selectivity between NQO1/CPR may indicate superior safety
profile [18,20-21]. To obtain the selectivity of the compounds, we incubated human
recombinant NQO1 or CPR with the test compounds and measured the reduction rate by the
two enzymes respectively. The reduction rate was recorded as µmol NADPH oxidized
min-1 •µmol-1 protein, and selectivity ratios were defined as the rate reduced by NQO1
divided by the rate reduced by CPR (Table 1).
The results of the reduction rates of the diverse quinone scaffolds catalyzed by
NQO1 or CPR were shown in Table 1, together with the reduction rate of β-lap, taken
as reference. Analysis of the data showed that most of these compounds tested could
be reduced by NQO1 and/or CPR. It was observed that compounds 12b-e (905-1623
µmol NADPH/min/µmol NQO1), 12f-i (1005-1295 µmol NADPH/min/µmol NQO1)
with substituted methyl groups at A ring showed better NQO1 catalytic efficiency
than corresponding parent compounds 12a (765 µmol NADPH/min/µmol NQO1) or
12j (745 µmol NADPH/min/µmol NQO1), indicated methyl-substitution was
favorable for NQO1 substrates. Moreover, the methyl-substituted compounds (12b-i)
(selectivity ratio = 1.17-6.37) also showed better selectivity and specificity toward
8

ACCEPTED MANUSCRIPT
NQO1 than compounds 12a (selectivity ratio = 0.63) and 12j (selectivity ratio = 0.54).
It should be noticed that that though compound 12k exhibited high reduction rate
(1446 µmol NADPH/min/µmol NQO1) by NQO1, it could also be reduced by CPR
easily (1221 µmol NADPH/min/µmol CPR). Generally, introduction of methyl group
on the A-ring increased the specificity toward NQO1 and substituents on the C-ring
exerted very little influence on the selectivity. The ability of NQO1 to process
compounds 17a-g in vitro has been assessed in previous work [25], we determined
their reduction rates by CPR and calculated selectivity ratio (NQO1/CPR) in this
work. As illustrated in Table 1, 17a-d showed poor selectivity toward NQO1 with
selectivity ratio at the range of 0.52-1.10, indicated that they could not be superior
scaffold for NQO1 specific substrates. Compounds (17e-g) with methoxy group at the
A ring showed better catalytic efficiency and specificity than 17a, indicating that
methoxy-substitution was favorable for selectivity toward NQO1 over CPR. As for
the para-quinone scaffolds (18, 19, 27 and 29) (selectivity ratio = 0.72-0.91), they
possessed worse selectivity toward NQO1 than most of the ortho-quinone
compounds, indicated that they were also not suitable scaffolds as specific substrates
of NQO1. Overall, among these compounds, two scaffolds (eg., 12b and 12g) showed
much better selectivity toward NQO1 than β-lap (selectivity ratio =1.36). Especially
for compound 12b, which had two methyl groups at A ring, afforded the greatest
selectivity to NQO1 (selectivity ratio = 6.37). Quinone scaffolds with good selectivity
(selectivity ratio > 1.2) were selected for further evaluation.
Table 1
The reduction rates of the diverse quinone scaffolds catalyzed by NQO1 or CPR and the
electrochemical reduction potentials versus Ferrocenea
reduction rates (µmol
selectivity
ratio
Eredox
(V) vs
Fc
cpd
R1
R2
R3
R4
R5
NADPH/min/µmol enzyme)
NQO1 CPR
(NQO1/CPR)
9

ACCEPTED MANUSCRIPT
12a
Me
H
H
H
Me
H
H
Me
H
H
H
765 ± 22
905 ± 53
1201 ± 92
142 ± 25
672 ± 30
835 ± 50
891 ± 75
818 ± 19
400 ± 42
796 ± 74
1014 ± 52
1382 ± 26
1221 ± 29
0.63
6.37
1.38
1.25
1.82
1.28
2.51
1.63
1.17
0.54
1.18
-1.03
-1.09
-1.07
-1.05
-1.04
−
12bb Me
12cb Me Me
Me
H
927 ± 11
12db Me
12eb Me
H
H
H
H
H
H
H
H
H
Me
H
1040 ± 38
1623 ± 112
1046 ± 11
1005 ± 81
1295 ± 40
1182 ± 47
745 ± 18
Me
Me
Me
Me
H
H
12fb
12gb
12hb
12ib
12jb
12k
Et
H
Me
Me
H
H
H
−
Et
Et
Et
H
H
−
Me
H
H
−
H
H
−
H
H
H
1446 ± 111
-0.98
17a
17b
17c
17d
17e
17f
Me Me
H
Me
H
−
−
−
−
−
-
-
-
1000 ± 50
753 ± 68
648 ± 47
719 ± 38
1247 ± 27
1422 ± 22
1300 ± 79
1437 ± 61
1427 ± 134
905 ± 27
659 ± 45
782 ± 19
812 ± 31
764 ± 33
0.70
0.52
0.71
1.10
1.59
1.75
1.70
-1.01
-
H
H
H
H
H
−
−
−
Me
H
−
−
Me Me
Me Me
H
C7-OMe
C8-OMe
C9-OMe
-1.09
-1.13
-1.10
H
17g Me Me
H
18
719 ± 38
730 ± 41
1010 ± 10
1011 ± 38
1422 ± 49
0.72
0.72
0.81
-1.22
-1.19
-0.90
19
27
1152 ± 298
29
774 ± 5
852 ± 19
923 ± 49
0.91
1.36
-1.22
-1.02
β-lap
−
1258 ± 46
a Eredox values calculated as (Epc + Epa)/2 are average values from voltammograms recorded at potential sweep rate
of 50 mV/s. Epc = cathodic peak potential; Epa = anodic peak potential.
3.3. Electrochemical studies of the quinones.
10

ACCEPTED MANUSCRIPT
Electrochemistry is the standard method for studying redox reactions, and many
works were replete with electrochemical techniques as applied to biology [29].
Several studies have investigated and reported on the electrochemical behavior of
naphthoquinones [30-31]. The electrochemical parameters always correlated with
biological activities [29]. In this work, the electrochemistry studies were performed to
explain the electrochemical behavior of the selective reduction rates for the substrates
of NQO1 to CPR with different scaffolds.
Representative compounds were selected to encompass each of the structural
classes. Dried dimethyl sulfoxide (DMSO) was used as the solvent for all compounds,
and tetrabutylammonium hexafluorophosphate (Bu4NPF6) was used as the supporting
electrolyte. Cathodic and anodic peak potentials, Epc and Epa were measured
respectively at a potential sweep rate of 50 mV/s, and the midpoint of the peak
potentials was used to determine Eredox values, (Epc + Epa)/2. The Eredox values, with
reference to ferrocene (Fc 0/+), was shown in Table 1 and the cyclic voltammograms
(CV) of the representative compounds were shown in Fig. 3.
As illustrated in Table 1, these investigated naphthoquinones possessed Eredox
values in the range of -0.93 to -1.22 V. Compounds with electron-donating groups at
A ring or C ring (compound 12a-12e or compound 17e-17g) showed slightly more
negative Eredox values (-1.03 to -1.09 V or -1.09 to -1.13 V), compare to 12k (-0.98 V)
or 17a (-1.01V). In addition, as observed in Fig. 3, two cathodic peaks and
corresponding backward peaks are displayed for these selected compounds, revealing
these compounds exhibited redox systems with a quase-reversible nature．
In our original plan, electrochemical reduction potentials was employed to
correlate with the selective reduction rates of the naphthoquinones by NQO1 or CPR,
however, after analyzing the results, there is only very small association between
them. This finding along with our molecular modeling studies suggest that other
factors such as ligand-receptor interaction may be more important. However,
according to the electrochemical study of the most selective compound, we
hypothesized that the potential of substrates in appropriate range would possess good
selectivity towards NQO1 over CPR. And more specific work need to be done in
11

ACCEPTED MANUSCRIPT
order to prove it and we would focus on this aspect in the future.
Fig. 3. Cyclic voltammograms of the representative compounds 12b, 12c, 12d, 12e, 12k, 17e, 17f,
β-lap and the control Ferrocene. Glassy carbon electrode, versus Ag/AgCl, Cl-1, cathodic
direction, v = 100 mV• s-1.
3.4. In vitro cytotoxicity evaluation of the selected quinone scaffolds.
On the basis of the enzymatic assays and electrochemical studies, potent and
selective NQO1 substrates were selected for cellular assay. These selected scaffolds
were investigated for their potency against the human NSCLC cell line A549 and
human normal liver cell line L02. According to the results of our determination, A549
has robust expression of NQO1 with relatively low expression of CPR, which is
opposite of L02 (the result of the protein activity in each cell lysate was shown in Fig.
4). Both of the two cell lines were exposed to the tested quinones for 24 h, and cell
death was assessed at 72 h. In parallel with the enzymatic results, most of the selected
compounds (selectivity ratio > 1.2) showed potent cytotoxicity against A549 cells
12

ACCEPTED MANUSCRIPT
with IC50 values ranging from 2.9 to 12 µM. On the contrary, these compounds
possessed relative poor cytotoxicity against L02 cells (Table 2). Compounds with
better selectivity towards A549 were supposed to have a superior safety profile. The
results demonstrated that more selective NQO1 substrates in vitro possessed better
antitumor activity. It was worth noting that the most selective substrate for NQO1
(12b) was potent to the NQO1-rich A549 cell lines and weakly active to the CPR-rich
L02 cell lines, shifting the IC50 10.6-fold (the fold is the ratio of the IC50 of A549 to
IC50 of L02, and a higher ratio indicates more NQO1 specificity and safety). This
compound showed much more safety than the control β-lap (shifting the IC50
3.67-fold) in cellular assay. In an effort to obtain antitumor drugs with efficacy and
safety, compound 12b was also assessed versus various other NQO1-expressing
cancer cells (MCF-7, HepG2, HCT116, H1299, H522 and H1650), NQO1-deficient
cells (A549 cells pretreated with NQO1 inhibitor dicoumarol (DIC) and H596) (Table
3). Compound 12b treatment markedly induced cell death in NQO1-expressing cancer
cells (IC50 = 2.1~4.9 µM), while with weaker toxic effects towards NQO1-deficient
cell lines (A549 + DIC and H596, IC50 = 21 µM and 35 µM, respectively). The results
confirmed that compound 12b had a NQO1-expressing cancer-cell-selective killing
property.
Fig. 4. NQO1/CPR expression in A549 cancer cell and L02 normal live cell lines.
Table 2
Cytotoxicity of the representative compounds toward A549 (NQO1-rich cancer cells)
and L02 (CPR-rich normal liver cells).
13

ACCEPTED MANUSCRIPT
selectivity ratio
IC50 (L02)/ IC50
(A549)
cytotoxicity IC50 (µM)
selectivity ratio
(NQO1/CPR)
entry
cpd
A549
L02
6.37
1.38
1.25
1.82
1.28
2.51
1.63
1.59
1.75
1.70
1.36
1
2
12b
12c
12d
12e
12f
3.3 ± 0.7
2.9 ± 1.7
4.1 ± 1.6
8.5 ± 1.1
11 ± 4.7
35 ± 0.9
17 ± 2.8
6.3 ± 0.1
22 ± 2.1
19 ± 1.2`
9.0 ± 0.2
25 ± 11
30 ± 8.1
32 ± 4.0
19 ± 2.9
11 ± 1.3
10.6
5.86
1.54
2.59
1.73
2.90
8.06
2.50
5.51
1.73
3.67
3
4
5
6
12g
12h
17e
17f
3.1 ± 2.2
3.1 ± 1.2
12 ± 3.8
5.8 ± 1.2
11 ± 0.9
3.0 ± 0.2
7
8
9
10
11
17g
β-lap
Table 3
Cell death of NQO1-expressing cancer cells. (A549, MCF-7, HepG2, HCT116, H1299, H522 and
H1650), NQO1-deficient cells (A549 cells pretreated with NQO1 inhibitor dicoumarol, H596) and
normal hepatic cells (L02) treated with compound 12b.
Cytotoxicity IC50 (µM)
Cpds
A549
(+)a
MCF-7
(+)a
HepG2
(+)a
HCT116
(+)a
H1299
(+)a
L02
(-)b
A549 + H596
DIC (-)b (-)b
H522(+)a H1650(+)a
9.3±0.7
3.3±0.7 2.4±0.8 2.5±0.4
4.4±1.1
4.9±0.7 4.8±0.9
35±0.9 21±4.9 35±9.1
12b
a (+) represents the NQO1-rich cancer cells.
b (-) represents the NQO1-deficient cells.
The representative compound 12b was further evaluated for its therapeutic
window using A549 cancer cells and L02 normal liver cells as surrogates. Prolonged
exposures (more than 8 h) of cells to β-lap abrogated the therapeutic window of this
drug in that normal cells (L02) was significantly killed by β-lap (Fig. 5A), probably
through the one-electron oxidoreductions [16]. As mentioned above, 12b showed
better selectivity toward NQO1 as compared to β-lap. Consistently, longer prolonged
exposures (24 h) of cells to 12b still showed an efficient antitumor activity with a
wide therapeutic window (Fig. 5B).
14

ACCEPTED MANUSCRIPT
Fig. 5. Defining the therapeutic window of β-lap (A) and 12b (B). Relative survival was
monitored in A549 cancer cells or L02 normal liver cells treated with various 12b concentrations
(µM) for 2 h, 8 h, and 24 h. Graphed are means with standard error of duplicate experiments (n ≥
3).
3.5. Molecular modeling.
Utilizing the crystal structure of NQO1 (PDB code 2F1O), computational
docking study was performed to elucidate the possible binding mode of NQO1 with
β-lap and 12b (Fig. 6). Both of them were found to form π-stacking interactions with
the isoalloxazine ring of FAD which was similar to the parent substrate. As previously
reported [19], the two carbonyl groups in the β-lap could bind firmly to the Tyr126
and Tyr128 residues by hydrogen bonding interactions. Similar to β-lap, compound
12b could also form these interactions. In addition, Trp105, Phe106 and Phe178
formed a hydrophobic pocket of sufficient size to allow only small substituents to
interact favorable. The 7,8-dimethyl groups of 12b can form C-H…π interactions with
the adjacent Phe178 and Trp105 residues which was observed for enhancing
π-stacking interaction with surrounding residues, supporting 12b was excellent
substrate for NQO1 from the other hand. The docking study provided a rational
explanation for better NQO1 specificity.
15

ACCEPTED MANUSCRIPT
Fig. 6. Docked conformation of compounds 12b (A) and β-lap (B) into active site of NQO1
(distance represents hydride donor-acceptor distance). The interaction mode was obtained through
molecular docking (PDB code: 2F1O) and depicted using MOE 2013.08. The carbon atoms of the
compounds and the key residues in the active site of NQO1 were colored in purple and yellow,
respectively. The H-bonds were shown as blue dot lines, the C-H…π interactions were shown as
purple dot lines.
3.6. Determination of ROS production.
It is supposed that 12b exerted its antitumor activity by accumulating ROS
through
futile
redox
cycles
catalyzed
by
intracellular
NQO1.
2’,7’-dichlorodihydrofluorescein diacetate (DCFH-DA) is a cell-permeable,
well-established probe to detect intracellular production of ROS. When the probe was
diffused into cells, DCFH-DA is deacetylated by cellular esterases to nonfluorescent
2’,7’-dichlorodihydrofluorescein, which is rapidly converted to highly fluorescent
2’,7’-dichlorofluorescein upon reacting with ROS. The probe is a leading approach to
measuring highly reactive species that may be most likely to initiate cancer cell death
[9]. As depicted in Fig. 7, the effect of 12b on cellular ROS levels in cancer cells
(A549) and normal cells (L02) was detected through flow cytometry using the
redox-sensitive fluorescent probe DCFH-DA. After treating A549 cells with 12b for 1
h and 3 h, a marked increase in ROS levels was caused. As shown in Fig. 7A, 12b
generated excessive ROS which showed comparable ability with the control β-lap in
A549 cells but lower in DIC and N-acetyl cysteine (NAC, a typical ROS scavenger)
pretreated A549 cells. In contrast to the results in cancer cell A549, 12b did not cause
an increase in ROS level in normal cell L02 (Fig. 7B). However, the control β-lap
16

ACCEPTED MANUSCRIPT
increased ROS level obviously in the normal cell L02, which may be induced by the
one-electron oxidoreductase CPR, and explained the potential toxicity of β-lap after
prolonged treatment. This selective induction of ROS in cancer cells indicates that
12b-induced ROS elevation is a consequence of cell transformation. The results
indicated that 12b induced ROS production in a NQO1-dependent and specific
manner and showed selective towards cancer cells. The property of better selective
induction of ROS in cancer cells than β-lap was promised to have a superior safety
profile.
Fig. 7. Induction of ROS in A549 or L02 cells. (A) 12b-Induced ROS elevation and reversion by
DIC or NAC. A549 cells were treated with 12b, β-lap or DMSO for 1 h and 3 h. Cells were also
pretreated with 10 µM DIC or 3 mM NAC for 1 h, followed by 10 µM 12b for 3 h. (B) 12b (10
µM, 3 h) does not increase ROS levels in normal cells (L02). ROS levels were measured by flow
cytometry and shown by quantitative bar graph measured as the fold change over DMSO-treated
levels. All values are mean ± SD of three independent experiments.
3.6. Further evaluation of compounds 12b in vivo.
To further verify that compound 12b was an efficient antitumor chemotype, we
proceeded to evaluate its in vivo antitumor activity in A549 tumor xenografts mouse
model (Fig. 8). When the tumor grew to a volume of 100-150 mm3, the nude mice
bearing A549 tumor xenografts were randomly injected with vehicle, β-lap (ARQ
17

ACCEPTED MANUSCRIPT
501, 30 mg/kg) and compound 12b (15 mg/kg or 30 mg/kg) every other day for three
weeks. The changes in tumor size and body weight were monitored twice per week
over a 21-day period. Treatment with 15 mg/kg and 30 mg/kg 12b both significantly
inhibited the growth of implanted lung tumor (P < 0.01 vs the vehicle-treated control,
Fig. 8A). Compound 12b (30 mg/kg) demonstrated potent in vivo antitumor activity
with comparable tumor growth inhibition (67.45%) to β-lap (70.70%). More
importantly, there was no statistical difference in body weight change among these
groups of mice by treated with 12b compared to β-lap (Fig. 8B). Together, our data
clearly demonstrated the efficacious of the novel quinone scaffold 12b.
Fig. 8. Compound 12b suppressed the tumor growth in vivo in A549 tumor xenografts
nude model. A549 tumor-bearing mice were treated with vehicle, 12b (15 mg/kg and
30 mg/kg) and β-lap (30 mg/kg). The relative tumor volume (A) and body weight (B)
of the treated mice are shown.
4. Conclusion
To develop quinone-directed antitumor agents with better selectivity and more
safety properties, a key scaffold of 3,7,8-trimethylnaphtho[1,2-b]furan-4,5-dione (12b)
was discovered as a specific NQO1 substrate through structure-activity relationship
studies. Compound 12b afforded the greatest selectivity to NQO1 (selectivity ratio =
6.37), which is much better than the control β-lap (selectivity ratio = 1.36). The
potential therapeutic window of 12b seems superior to the control β-lap, making this
scaffold a potentially safer candidate for use as an anticancer drug. In addition, the
18

ACCEPTED MANUSCRIPT
highlighting compound 12b exhibited comparable anticancer activity compared to
β-lap both in vitro and in vivo, which might be a promising and novel scaffold for
further investigation.
5. Experimental Section
5.1. Chemistry
5.1.1. General experimental methods.
All chemicals and reagents and were purchased from commercial sources.
Organic solutions were concentrated in a rotary evaporator (Büchi Rotavapor) below
55 oC under reduced pressure. Silica gel thin-layer chromatography was performed on
precoated plates GF-254 and visualized under UV light. Melting points were
determined with a Melt-Temp II apparatus. 1H NMR and 13C NMR spectra were
recorded on a Bruker AV-300 instrument using deuterated solvents with
tetramethylsilane (TMS) as internal standard. Chemical shifts are given in ppm (δ).
The multiplicities are denoted as follows: s, singlet; d, doublet; t, triplet; q, quartet; m,
multiplet; br s, broad singlet. IR spectra were recorded on a Nicolet iS10 Avatar
FT-IR spectrometer using KBr film. ESI-mass and high resolution mass spectra
(HRMS) were recorded on a Water Q-Tofmicro mass spectrometer. Analytical results
are within 0.40 % of the theoretical values. Purity of the compounds was analyzed by
HPLC (Agilent Technologies 1260 Infinity) using 80:20 MeOH/H2O as the mobile
phase with a flow rate of 0.8 mL/min on a C18 column. All tested compounds
exhibited greater than 95% purity unless otherwise noted.
5.1.2. Synthesis
5.1.2.1. 3,7,8-Trimethylnaphtho[1,2-b]furan-4,5-dione (12b).
Step 1: 3-methyl-5-hydroxyphenol [1,2-b] furan. To a solution of 1,4-benzoquinone 6
(21.6 g, 200 mmol) in dichloromethane (200 mL) was added
o
(E)-4-(Prop-1-en-1-yl)morpholine 8b (25.4 g, 200 mmol) at 0 C. The reaction was
then kept for 12 h at room temperature. The black precipitate was filtered and washed
with dichloromethane. The solvent was distilled and the product 9b was ready for the
19

ACCEPTED MANUSCRIPT
next step without the further purification. The crude product 9b was dissolved in THF
(50 mL) and 150 mL 10% HCl was added. The reaction mixture was refluxed for 2 h,
and then cooled and extracted be ethyl acetate (3 × 100 mL). The organic layers were
combined, dried over anhydrous Na2SO4 and concentrated to give a light purple solid,
which was further purified by recrystallization from dichloromethane to give the titled
compound as a white solid 10b (16 g, 54% for two steps). m.p. 140-142 oC. 1H NMR
(300 MHz, CDCl3) δ: 7.39 (d, J = 1.2 Hz, 1H), 7.30 (d, J = 8.7 Hz, 1H), 6.95 (d, J =
2.4 Hz, 1H), 6.81 (dd, J = 2.6 Hz, 2.6 Hz, 1H), 4.73 (s, 1H), 2.20 (s, 3H). m/z
(EI-MS): 148 [M]+.
Step 2: 3,7,8-Trimethylnaphtho[1,2-b]furan-4,5-dione (12b). To a solution of 10b (2
g, 14 mmol) in acetone/H2O (100 mL, v/v = 1:1) added 2-iodoxybenzoic acid (IBX, 5
g, 19 mmol). The reaction mixture was stirred at room temperature for 1 h. After that,
the precipitate was filtered and the filtrate was then extracted by toluene (3 × 100
mL). The solvent was partly distilled (about 50 mL left), then diisopropenyl and
2,3-dicyano-5,6-dichlorobenzoquinone (DDQ) was added to the solution. The
reaction was refluxed at 110 oC for 3 h. Then, the solvent was vacuum-evaporated and
the residue was redissolved in ethyl acetate. The organic layer was washed with
saturated NaHCO3 aqueous and brine, dried over anhydrous Na2SO4, filtrated and
concentrated to afford the crude product, which was purified by column
chromatography (eluent: dichloromethane) to afford 12b as a red solid (500 mg,
o
1
15%). m.p. 180-182 C. H NMR (300 MHz, CDCl3) δ: 7.84 (s, 1H), 7.46 (s, 1H),
13
7.24 (s, 1H), 2.38 (s, 3H), 2.34 (s, 3H), 2.29 (s, 3H). C NMR (75 MHz, CDCl3) δ:
180.07, 175.20, 160.52, 144.98, 140.57, 138.70, 130.99, 126.18, 125.81, 122.83,
120.93, 119.96, 19.86, 19.21, 8.35. IR (v, cm-1): 3145, 1671, 1593, 1539, 1030.
ESI-HRMS m/z [M + H]+ calculated for C15H13O3: 241.0859, found: 241.0855.
5.1.2.2. 3,6,9-Trimethylnaphtho[1,2-b]furan-4,5-dione (12c).
Following the synthetic method of compound 12b, the product 12c was obtained as
1
red solid: yield 22%. m.p. 203-204 oC. H NMR (300 MHz, CDCl3) δ: 8.09 (d, 1H, J
= 7.8 Hz), 7.69-7.62 (m, 2H), 7.47 (t, 1H, J = 9.0 Hz), 2.31 (s, 3H). IR (ν, cm-1): 3416,
20

ACCEPTED MANUSCRIPT
2360, 2342, 1670, 1507, 1242. ESI-HRMS m/z [M + H]+ calculated for C15H13O3:
241.0859, found: 241.0854.
5.1.2.3. 3,8-Dimethylnaphtho[1,2-b]furan-4,5-dione (12d).
Following the synthetic method of compound 12b, the product 12d was obtained as
red solid: yield 13%. m.p. 192-193 oC. 1H NMR (300 MHz, CDCl3) δ: 7.96 (d, J = 8.1
Hz, 1H), 7.50 (s, 1H), 7.23 (s, 2H), 2.46 (s, 3H), 2.29 (s, 3H). IR (ν, cm-1): 3447,
2360, 2342, 1673, 1223. ESI-HRMS m/z [M + H]+ calculated for C14H11O3:
227.0703, found: 227.0708.
5.1.2.4. 3,7-Dimethylnaphtho[1,2-b]furan-4,5-dione (12e).
Following the synthetic method of compound 12b, the product 12e was obtained as
o
1
red solid: yield 16%. m.p. 216-217 C. H NMR (300 MHz, CDCl3) δ: 7.90 (s, 1H),
7.60 (d, J = 7.8 Hz, 1H), 7.46 (d, J = 7.1 Hz, 1H), 7.26 (d, J = 1.3 Hz, 1H), 2.44 (s,
3H), 2.31 (d, J = 1.2 Hz, 3H). IR (ν, cm-1): 3416, 2963, 2926, 1675, 1091, 804.
ESI-HRMS m/z [M + H]+ calculated for C14H11O3: 227.0703, found: 227.0701.
5.1.2.5. 3-Ethyl-7,8-dimethylnaphtho[1,2-b]furan-4,5-dione (12f).
Following the synthetic method of compound 12b, the product 12f was obtained as
o
1
orange solid: yield 32%. m.p. 229-231 C. H NMR (300 MHz, CDCl3) δ: 7.85 (s,
1H), 7.47 (s, 1H), 7.25 (s, 1H), 2.78-2.70 (m, 2H), 2.39 (s, 3H), 2.34 (s, 3H), 1.27 (t, J
= 7.4 Hz, 3H). IR (ν, cm-1): 3421, 2360, 2342, 1665, 1231. ESI-HRMS m/z [M + H]+
calculated for C16H15O3: 255.1016, found: 255.1011.
5.1.2.6. 7,8-Dimethylnaphtho[1,2-b]furan-4,5-dione (12g). Following the synthetic
method of compound 12b, the product 12g was obtained as red solid: yield 28%. m.p.
205-206 oC. 1H NMR (300 MHz, CDCl3) δ: 7.87 (s, 1H), 7.50-7.48 (m, 2H), 6.86 (d, J
= 1.7 Hz, 1H), 2.40 (s, 3H), 2.35 (s, 3H). IR (ν, cm-1): 3420, 2923, 2853, 1672, 1109.
ESI-HRMS m/z [M + Na]+ calculated for C14H10O3Na: 249.0522, found: 249.0527.
21

ACCEPTED MANUSCRIPT
5.1.2.7. 3-Ethyl-7-methylnaphtho[1,2-b]furan-4,5-dione (12h). Following the
synthetic method of compound 12b, the product 12h was obtained as red solid: yield
o
1
18%. m.p. 208-210 C. H NMR (300 MHz, CDCl3) δ: 7.88 (s, 1H), 7.58 (d, J = 7.7
Hz, 1H), 7.44 (d, J = 8.4 Hz, 1H), 7.24 (s, 1H), 2.72 (q, J = 7.4 Hz, 2H), 2.42 (s, 3H),
1.25 (t, J = 7.4 Hz, 3H). IR (ν, cm-1): 3421, 2360, 2346, 1665, 1231. ESI-HRMS m/z
[M + H]+ calculated for C15H13O3: 241.0859, found: 241.0863.
5.1.2.8. 3-Ethyl-8-methylnaphtho[1,2-b]furan-4,5-dione (12i). Following the synthetic
method of compound 12b, the product 12i was obtained as red solid: yield 21%. m.p.
o
1
195-196 C. H NMR (300 MHz, CDCl3) δ: 7.96 (d, J = 7.8 Hz, 1H), 7.50 (s, 1H),
7.32 (s, 1H), 7.23 (s, 1H), 2.72 (q, J = 7.4 Hz, 2H), 2.42 (s, 3H), 1.25 (t, J = 7.4 Hz,
3H). IR (ν, cm-1): 3421, 2970, 2361, 2342, 1670, 1093. ESI-HRMS m/z [M + H]+
calculated for C15H13O3: 241.0859, found: 241.0856.
5.1.2.9. 3-Ethylnaphtho[1,2-b]furan-4,5-dione (12j). Following the synthetic method
of compound 12b, the product 12j was obtained as red solid: yield 26%. m.p. 220-222
1
oC. H NMR (300 MHz, CDCl3) δ: 8.06 (d, J = 7.5 Hz, 1H), 7.72-7.62 (m, 2H),
7.48-7.43 (m, 1H), 7.28 (s, 1H), 2.73 (q, J = 7.4 Hz, 2H), 1.26 (t, J = 7.4 Hz, 3H). IR
(ν, cm-1): 3415, 2971, 1677, 1215, 1088. ESI-HRMS m/z [M + H]+ calculated for
C14H11O3: 227.0703, found: 227.0707.
5.1.2.10. Naphtho[1,2-b]furan-4,5-dione (12k). Following the synthetic method of
compound 12b, the product 12k was obtained as red solid: yield 45%. m.p. 186-187
1
oC. H NMR (300 MHz, CDCl3) δ: 8.09 (d, J = 7.2 Hz, 1H), 7.72-7.67 (m, 2H),
7.52-7.48 (m, 2H), 6.88 (d, J = 2.1 Hz, 1H). IR (ν, cm-1): 3415, 2971, 1677, 1215,
1088. ESI-HRMS m/z [M + H]+ calculated for C12H7O3:199.0390, found: 199.0395.
5.1.2.11. 3-Methylnaphtho[1,2-b]furan-4,5-dione (12a). The synthetic method for
compound 12a was reported as our previous report [31], the product was obtained as
1
red solid: yield 60%. m.p. 170-172 oC. H NMR (300 MHz, CDCl3) δ: 8.09 (d, 1H, J
= 7.8 Hz), 7.69-7.62 (m, 2H), 7.47 (t, 1H, J = 9.0 Hz), 2.31 (s, 3H). IR (ν, cm-1): 3134,
22

ACCEPTED MANUSCRIPT
1672, 1590, 1539, 1020. ESI-HRMS m/z [M+Na]+ calculated for C13H8O3Na:
235.0371, found: 235.0374.
5.1.2.12. 2,3-Dihydro-2,3,3-dimethylnaphtho[1,2-b]furan-4,5-dione (17a). To a
solution of ether (16) (50 mg, 0.2 mmol) in DMF (5 mL) was added the Lewis acid
o
NbCl5 (1 mmol). The mixture was stirred at 120 C for 0.5 h. Then the reaction
mixture was poured into cooled water (10 mL), extracted with ethyl acetate (3 × 10
mL), and washed with brine (20 mL). The organic layer was dried and concentrated,
and the pure product dunnione (17a) (46 mg, 92%) was obtained as a red solid
o
1
without any need for further purification. m.p.: 93-95 C. H NMR (300 MHz,
CDCl3): δ = 8.06 (d, J = 7.5 Hz, 1H), 7.65 (d, J = 3.9 Hz, 2H), 7.56-7.59 (m, 1H),
13
4.68 (q, J = 9Hz, 1H), 1.48 (d, J = 6.6Hz, 3H), 1.46 (s, 3H), 1.28 (s, 3H) ppm. C
NMR (75 MHz, CDCl3): δ = 181.0, 175.0, 167.9, 133.9, 131.1, 130.2, 128.7, 127.5,
123.9, 122.9, 92.4, 43.7, 25.3, 19.9, 14.1 ppm. IR (KBr) ν 3450, 1638, 1612, 1590,
1572, 1401, 1381 cm-1. ESI-HRMS calcd for C15H14O3Na [M + Na]+ 265.0835,
found 265.0831.
5.1.2.13. 2,3-Dihydro-2,2-dimethylnaphtho[1,2-b]furan-4,5-dione (17b). To a
solution of ether (16) (46 mg, 0.2 mmol) in DMF (5 mL) was added the Lewis acid
NbCl5 (1 mmol). The mixture was stirred at 120 oC for 2 h. Then the reaction mixture
was poured into cooled water (10 mL), extracted with ethyl acetate (3 × 10 mL), and
washed with brine (20 mL). The organic layer was dried and concentrated.
Purification via silica column chromatograph (petroleum ether/ethyl acetate = 20/1,
o
1
v/v) afford 17b as red solid (36 mg, 78% yield). m.p.: 169-171 C. H NMR (300
MHz, CDCl3): δ = 8.08 (d, J =7.5 Hz, 1H), 7.64 (d, J = 3.9 Hz, 2H), 7.56-7.59 (m,
13
1H), 2.95 (s, 2H), 1.61 (s, 6H) ppm. C NMR (75 MHz, CDCl3): δ = 180.8, 175.2,
168.2, 133.9, 131.3, 130.4, 128.8, 127.5, 124.1, 114.5, 93.2, 38.8, 27.9 (2 × CH3)
ppm. IR (KBr) ν 3453, 1693, 1643, 1610, 1567, 1403, 1278 cm-1. ESI-HRMS: calcd
for C14H12O3Na [M + Na]+ 251.0679, found 251.0676.
23

ACCEPTED MANUSCRIPT
5.1.2.14. 2,3-Dihydro-2-methylnaphthol[1,2-b]furan-4,5-dione (17c). To a solution of
ether (16) (43 mg, 0.2 mmol) in DMF (5 mL) was added the Lewis acid NbCl5 (1
o
mmol). The mixture was stirred at 120 C for 2 h. Then the reaction mixture was
poured into cooled water (10 mL), extracted with ethyl acetate (3 × 10 mL), and
washed with brine (20 mL). The organic layer was dried and concentrated.
Purification via silica column chromatograph (petroleum ether/ethyl acetate = 20/1,
o
1
v/v) afford 17c as orange solid (33 mg, 78% yield). m.p.: 125-126 C. H NMR (300
MHz, CDCl3): δ = 8.08 (d, J =7.2 Hz, 1H), 7.65 (d, J = 3.6 Hz, 2H), 7.57-7.60 (m,
1H), 5.25-5.32 (m, 1H), 3.28 (dd, J = 9.9Hz, J = 9.6 Hz, 1H), 2.75 (dd, J =7.2 Hz, J
13
=6.8 Hz, 1H), 1.58 (d, J =5.7 Hz, 3H) ppm. C NMR (75 MHz, CDCl3): δ = 181.0,
175.2, 169.1, 133.9, 131.1, 130.4, 129.2, 127.5, 123.9, 115.5, 85.5, 33.2, 21.8 ppm. IR
(KBr) ν 3452, 1692, 1647, 1600, 1555, 1400, 1159 cm-1. ESI-HRMS: calcd for
C13H10O3Na [M + Na]+ 237.0522, found 237.0518.
5.1.2.15. 2,3-Dihydro-2,3-dimethylnaphthol[1,2-b]furan-4,5-dione (17d). To a
solution of ether (16) (46 mg, 0.2 mmol) in DMF (5 mL) was added the Lewis acid
NbCl5 (1 mmol). The mixture was stirred at 120 oC for 2 h. Then the reaction mixture
was poured into cooled water (10 mL), extracted with ethyl acetate (3 × 10 mL), and
washed with brine (20 mL). The organic layer was dried and concentrated.
Purification via silica column chromatograph (petroleum ether/ethyl acetate = 30/1,
o
v/v) afford 17d as orange solid (32 mg, 71% yield). m.p.: 137-138 C. 1H NMR (300
MHz, CDCl3): δ = 8.10 (d, J = 4.5 Hz, 1H), 7.66 (d, J = 3.9 Hz, 2H), 7.56-7.60 (m,
1H), 5.19 (dq, J = 2.4 Hz, J =8.8 Hz, 1H), 3.55 (dq, J = 2.1 Hz, 8.8 Hz), 1.56 (d, J =
13
6.6 Hz, 3H), 1.28 (d, J = 6.6 Hz, 3H) ppm. C NMR (75 MHz, CDCl3): δ = 181.0,
175.0, 167.7, 133.9, 131.3, 130.4, 128.8, 127.5, 124.1, 115.5, 86.7, 36.6, 14.8, 12.5
ppm. IR (KBr) ν 3431, 1680, 1644, 1598, 1545, 1360, 1100 cm-1. ESI-HRMS: calcd
for C14H12O3Na [M + Na]+ 251.0679, found 251.0675.
5.1.2.16. 7-Methoxy-2,3-dihydro-2,3,3-dimethylnaphtho[1,2-b]furan-4,5-dione (17e).
Following the synthetic method of compound 17a, the product 17e was obtained as
orange solid: yield 56%. m.p.: 180-181 oC. 1H NMR (300 MHz, CDCl3): δ = 7.58 (d,
24

ACCEPTED MANUSCRIPT
J = 2.7 Hz, 1H), 7.56 (s, 1H), 7.11 (dd, J = 2.7 Hz, J = 2.7 Hz, 1H), 4.65 (q, J = 6.9
13
Hz, 1H), 3.92 (s, 3H), 1.48 (d, J = 6.0 Hz, 3H), 1.45 (s, 3H), 1.27 (s, 3H) ppm. C
NMR (75 MHz, CDCl3): δ = 181.2, 174.9, 168.5, 162.2, 132.2, 125.7, 120.8, 120.1,
119.6, 113.7, 92.4, 55.4, 43.4, 25.3, 19.9, 14.1 ppm. IR (KBr) ν 3393, 1670, 1654,
1598, 1548 cm-1. HRMS (ESI+): calcd for C16H16O4Na [M + Na]+ 295.0941, found
295.0938.
5.1.2.17. 8-Methoxy-2,3-dihydro-2,3,3-dimethylnaphtho[1,2-b]furan-4,5-dione (17f).
Following the synthetic method of compound 17a, the product 17f was obtained as
red solid: yield 64%. m.p.: 175-177 oC. 1H NMR (300 MHz, CDCl3): δ = 8.03 (d, J =
8.4 Hz, 1H), 7.11 (d, J = 8.4 Hz, 1H), 6.98 (dd, J = 2.4 Hz, J = 2.4 Hz, 1H), 4.64 (q, J
= 6.6 Hz, 1H), 3.93 (s, 3H), 1.46 (d, J = 6.6 Hz, 3H), 1.44 (s, 3H), 1.28 (s, 3H) ppm.
IR (KBr) ν 3411, 1698, 1604, 1588, 1522 cm-1. HRMS (ESI+): calcd for C16H16O4Na
[M + Na]+ 295.0941, found 295.0934.
5.1.2.18. 9-Methoxy-2,3-dihydro-2,3,3-dimethylnaphtho[1,2-b]furan-4,5-dione (17g).
Following the synthetic method of compound 17a, the product 17g was obtained as
red solid: yield 50%. m.p.: 186-187 oC. 1H NMR (300 MHz, CDCl3): δ = 7.74 (d, J =
7.5 Hz, 1H), 7.50 (t, J = 8.4 Hz, 1H), 7.22 (d, J = 8.4 Hz, 1H), 4.64 (q, J = 6.6 Hz,
1H), 3.95 (s, 3H), 1.47 (d, J = 6.6 Hz, 3H), 1.44 (s, 3H), 1.25 (s, 3H) ppm. IR (KBr) ν
3396, 1700, 1675, 1600, 1544 cm-1. HRMS (ESI+): calcd for C16H16O4Na [M + Na]+
295.0941, found 295.0930.
5.1.2.19. 2,2-dimethyl-3,4-dihydro-2H-benzo[g]chromene-5,10-dione (18). To a
solution of lapachol (15) (50 mg, 0.2 mmol) in DCM (5 mL) was added the diluted
o
hydrochloric acid. The mixture was stirred at 45 C for 1 h and then the reaction
mixture was poured into cooled water (10 mL), extracted with ethyl acetate (3 × 10
mL) and washed with brine (10 mL). The organic layer was combined, dried over
sodium sulfate, and concentrated under reduced pressure to give the pure product
α-lapachone (18) (45 mg, 90%) without any need for further purification. mp 116–118
oC; 1H NMR (300 MHz, CDCl3) δ: 8.06 (m, 2H), 7.68 (m, 2H), 2.62 (t, J = 6.6 Hz,
25

ACCEPTED MANUSCRIPT
13
2H), 1.82 (t, J = 6.5 Hz, 2H), 1.44 (s, 6H). C NMR (75 MHz, CDCl3) δ: 183.9,
179.5, 154.1, 133.3, 132.4, 131.6, 130.7, 125.8, 125.5, 119.6, 77.6, 30.9, 26.0, 16.2.
HRMS-ESI m/z [M+H]+ calculated for C15H15O3: 243.1016, found: 243.1013.
5.1.2.20. 2,2,6,9-Tetramethyl-2H-pyrano[2,3-g]quinoline-5,7,10(6H)-trione (27). A
50 mL round-bottomed flask with a magnetic stir bar was charged with substituted
1,4-naphthoquinone (26) (0.50 mmol), Pd(OAc)2 (11.2 mg, 10.0 mol %), Cu(OAc)2
(199.6 mg, 2.0 equiv), NaOAc (82.0 mg, 2.0 equiv), 28 (2.5 mmol, 5.0 equiv), and
o
THF (15 mL) under air. The resulting mixture was heated at 45 C for 3 h and then
cooled to ambient temperature. The mixture was filtered through a celite pad, and
then washed with cooled water (20 mL), extracted with ethyl acetate (3 × 20 mL), and
washed with brine (40 mL), dried over anhydrous Na2SO4 and concentrated. The
residue was purified by column chromatography on silica gel (petroleum ether/ethyl
acetate = 5/1, v/v. Rf = 0.16) afford 27 as a dark red solid (140.0 mg, 45% yield). m.p.
o
1
134-135 C. H NMR (300 MHz, CDCl3) δ 6.60 (s, 1H), 6.52 (d, J = 9.9 Hz, 1H),
5.68 (d, J = 10.0 Hz, 1H), 3.87 (s, 3H), 2.57 (s, 3H), 1.56 (s, 6H) ppm. 13C NMR (75
MHz, CDCl3): δ = 178.3, 161.6, 150.8, 148.8, 141.8, 129.4, 125.1, 123.9, 115.3,
114.6, 114.0, 80.8, 34.1, 27.8, 22.3 ppm. ESI-HRMS: calcd for C16H16NO4 [M+H]+
286.1074, found 286.1076.
5.1.2.21.
2,2-Dimethyl-2H-benzo[g]chromene-5,10-dione
(α-xiloidone) (29).
Following the synthetic method of compound 27, the product 29 was obtained as
orange solid: yield 82%. m.p. 142-143 oC. 1H NMR (300 MHz, CDCl3) δ 8.10 (dd, J1
= 6.4 Hz, J2 = 2.5 Hz, 2H), 7.73-7.68 (m, 2H), 6.67 (d, J = 10.0 Hz, 1H), 5.74 (d, J =
10.0 Hz, 1H), 1.57 (s, 6H) ppm. 13C NMR (75 MHz, CDCl3): δ = 181.3, 179.3, 151.9,
133.4, 132.7, 130.9, 130.9, 130.4, 125.7, 117.3, 114.9, 79.9, 27.9 ppm. HRMS (ESI+):
calcd for C15H13O3 [M+H]+ 241.0859, found 241.0855.
5.2. Biology
5.2.1. In vitro enzyme reduction assay.
26

ACCEPTED MANUSCRIPT
Compounds were monitored as NQO1 or CPR substrates using UV-visible
spectroscopy by adapting a previous assay to well plate reader and 96-well quartz
plate [18]. The reduction rates were determined using an NADPH recycling assay and
the enzymes (NQO1 and CPR purchased from Sigma), in which NADPH oxidation to
NADP+ was monitored by absorbance (A340 nm) on a Varioskan Flash (Thermo,
Waltham, MA). The experiments were performed on three different days with at least
three replicates of each condition performed. The assay solution consisted of 0.1 M
phosphate buffer at pH 7.4 containing 0.007% of bovine serum albumin (BSA), 100
µL of appropriate concentration of compounds, 50 µL of 400 µM NADPH solutions,
and 50 µL of enzymes solutions (150 µM NQO1 and 200 µM one-electron
oxidoreductases). The total assay volume was 200 µL. All stock solutions were
prepared by dissolving each compound in DMSO (100 µL) with subsequent dilution
to 10 mL with 0.007% BSA solution. Once the 96-well plate was filled with the assay
solutions, except the NADPH, it was placed into the incubator at 37 oC and left to sit
for 3 min before the measurements. The enzyme reaction was initiated by automated
dispensing of the NADPH solution into the wells, and the absorbance change at 340
o
nm was measured at 2 s intervals for 5 min at room temperature (22-25 C). The
linear portion of the absorbance vs time graphs (the first 20 s to 1 min) were fitted and
the slops were calculated (velocity). NADPH oxidation rates were compared with
reactions lacking compound. Initial velocities were calculated and data expressed as
µmol NADPH oxidized/min/µmol protein. Plots of average velocity versus
compounds concentrations were used to obtain apparent Km and Vmax values from
Michaelis−Menten curves generated using Graphpad Prism 6.
5.2.2. Cell viability assay.
A549 and L02 cells were cultured in DEME (GiBco, USA). Both media
contained 10% fetal bovine serum (FBS, GiBco, USA). Cells were plated in 96-well
microtiter plates at a density of 10,000 cells/mL and incubated in a humidified
o
atmosphere with 5% CO2 at 37 C for 16 h. Test compounds were added onto
triplicate wells with different concentrations and 0.1% DMSO for the control. After
27

ACCEPTED MANUSCRIPT
they had been applied in medium for 24 h, removed, and replaced with fresh medium,
and
the
plates
were
incubated
for
72
h.
MTT
(3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) solution (5 mg/mL)
was added and the cells were incubated for another 4 h. The solutions were removed
carefully by aspiration, the formazan was dissolved in 100 µL of DMSO. The
absorbance (OD) was read on a plate at 570 nm. The concentration causing 50%
inhibition of cell growth (IC50) was determined by the Logit method.48 The data was
analyzed using Graphpad Prism 6.
5.2.3. Electrochemistry.
Cyclic voltammetry (CV) for representative compounds was conducted using a
model CS120 Electrochemical Analyzer (CH Instruments) equipped with a
conventional three-electrode system. The system was composed of a platinum counter
electrode as the auxiliary electrode, a Ag/AgCl as the reference electrode and a glossy
carbon (GC) rod as the working electrode. The reference electrode was separated from
the main compartment of the system by a salt bridge containing DMF solvent and
Bu4NBF4 supporting electrolyte. The reference and the salt bridge were calibrated by
voltammetry relative to Eredox for Ferrocene (0/+) couple in DMF/Bu4NBF4, to allow
the measured Eredox values for the naphthoquinones to be quoted relative to Ferrocene
(0/+). All of the compounds were run at concentration of 1 mM in DMF and all
samples were purged with nitrogen prior to use and kept under a continuous flow of
nitrogen during the course of the experiments. All data were recorded at a potential
range between 0.00 and -2.00 V and at potential sweep rates of 50 mV/s. All
measurements were performed at 25 oC.
5.2.4. In vivo antitumor activity.
Log-phase A549 cells were harvested, washed 3 times with PBS, followed by
resuspension at a concentration of 5 × 106 per mL. A total of 100 µL of cell
suspension was injected into the flanks of athymic nude mice (7-8 weeks). Tumor
sizes were regularly measured using calipers, and volumes were calculated using the
following formula: volume (mm3) = length × width × width/2. After the tumors had
grown to 100-150 mm3, all the mice were randomized into four groups (five mice for
28

ACCEPTED MANUSCRIPT
each group) in two independent experiments and dosed with 12b (15 or 30 mg/kg),
β-lap (ARQ 501, 30 mg/kg), or vehicle. Compound 12b was dissolved in HPβCD, and
ARQ501 was dissolved in normal saline. All agents were administered every other
day for three weeks through tail vein injection, and tumor growth was monitored and
measured every day. After three weeks, mice were euthanized and the average tumor
wet weights were calculated. Animal care and experiments were conducted in
accordance with the protocols approved by the Animal Ethics Committees of the
Institute of Materia Medical, Chinese Academy of Medical Sciences and China
Pharmaceutical University, following the Institutional Animal Care and Use
Committee guidelines.
5.2.5. Measurement of intracellular ROS levels.
Intracellular ROS was analyzed from the transformation of nonfluorescent
DCFH-DA to its fluorescent derivative. Fluorescence intensity was detected at 535
nm (with 488 nm excitation). Drug-treated cells were incubated with 10 µmol•L-1 of
DCFH-DA for 30 min at 37 oC, washed twice with PBS and immediately analyzed in
a FACScan flow cytometer. Experiments are representative of three separate
experiments.
5.2.6. Statistical analysis.
Results for continuous variables were presented as the mean ± standard error.
Two-group differences in continuous variables were assessed by the unpaired T-test.
P-values are two-tailed with confidence intervals of 95%. Statistical analysis was
performed by comparing treated samples with untreated controls. All statistical tests
were conducted using Prism software version 6.0 (GraphPad, San Diego, CA, USA).
5.3. Molecular modeling.
A docking study was performed using the crystal structure of the human NQO1
complex with dicoumarol (PDB code: 2F1O and resolution 2.75 Å) and the structure
was editor according to provide a monomer of the protein and protonated use GOLD
29